Scenario execution, cycles, and phases
is a transactional system, similar to relational databases. Each execution starts with the initialization phase, continues with at least one cycle composed of the operation and commit/rollback phases, and ends with the finalization phase.
During the initialization phase, all necessary connections are created. They are also checked if each module is capable of performing their intended operation(s).
Each cycle represents an undividable unit of work composed of a series of operations. It is possible to set the maximum number of cycles in the settings. The default number is 1.
During the operation phase, reading and writing operations are performed.
- The reading operation consists of obtaining data from a service that will then be processed by other modules according to a predefined . For example, the Dropbox > Watch files module returns new bundles (files) created since the last execution.
- The writing operation consists of sending data to a given service for further processing. For example, the Dropbox > Upload a file module uploads a file to a Dropbox folder.
If the operation phase is successful for all modules, the commit phase begins, during which all operations performed by the modules are committed. This means that Make sends information to all the services involved in the operation phase about its success.
If an error occurs during the operation or commit phase for any module, the phase is aborted and the rollback phase is started, making all operations during the given cycle void. Some modules do not support rollback and operations performed by these modules cannot be taken back.
Modules that support rollback and marked with the ACID tag. Modules without this tag do not support rollback and cannot be reverted back to their initial state in case of an error in other modules.
During the finalization phase, open connections (e.g. FTP connections, database connections, etc.) are closed and the is completed.
Transfer of bundles between databases
The following example shows how to connect three ACID modules. The aim of the is to get new rows from a MySQL database, insert (transfer) them into an MSSQL database, and then insert the IDs of the rows from the MSSQL database into a PostgreSQL database.
When the starts, the initialization phase is performed first. verifies connections to the MySQL, MSSQL and PostgreSQL databases one at a time. If everything goes well and the connections are successful, moves on to the operation phase. If an error occurs, the finalization phase starts instead of the operation phase and the is terminated.
If there isn't an error, the operation phase begins. A preset procedure selects (reads) the table rows (bundles) from MySQL. Those rows are then passed to the next module that writes them to a selected table in the MSSQL database. If everything is in order, the last PostgresSQL procedure is called to insert the row IDs returned by the preceding module into the table.
If the operation phase is completed successfully, the commit phase begins. calls the SQL COMMIT command for each database and the write operations are committed.
However, if the operation or commit phase fails due to an error, (e.g. connection failure), calls for a rollback. During the rollback phase, goes through all modules and executes the SQL ROLLBACK command for each module to revert each database back to its initial state.
Finally, during the finalization phase, each module closes its connection to the database.